Summary of reinforcement learning model fits
| Condition | α | β | WA | Hybrid AQ pseudo-R2 (AIC) | Q pseudo-R2 (AIC) |
|---|---|---|---|---|---|
| SHAM | 0.30 (±0.12) | 0.31 (±0.11) | 0.47 (±0.14) | 0.23 (369.7) | 0.20 (361.3) |
| STIM+ | 0.38 (±0.11) | 0.44 (±0.11) | 0.71 (±0.12) | 0.14 (404.7) | 0.07 (412.8) |
Means (±SEM) are shown for best-fitting AQ model parameter values (columns 2–4) associated with the STIM+ and SHAM pairs during Stage 2. We report pseudo-R2 and AIC goodness-of-fit measures for the three-parameter hybrid AQ model (column 5) and the two-parameter Q model (column 6) for each condition (see Materials and Methods).